Abstract:Fusion of multi-source information is one of the primary methods to alleviate data sparsity in recommender systems (RS). Hypergraphs have shown remarkable capabilities in dealing with the diversity of multi-source information, especially in modeling high-order user-item relationships. However, most hypergraph research focuses on constructing hypergraphs using a particular type of hyperedge, which might not fully capture the high-order implicit associations among heterogeneous information sources. Furthermore, existing Hypergraph Attention Networks (Hyper-GAT), mainly emphasizes on information propagation between nodes and hyperedges, insufficiently exploring density information within complex hyperedges. Moreover, when hypergraphs combine data from multiple heterogeneous graphs, redundant information from different viewpoints can reduce the effectiveness of multivariate modeling hyperedges. Thus, we propose a recommendation algorithm based on multi-source heterogeneous hypergraphs and contrastive learning (MHCLR), which improves recommendation accuracy through multi-source information fusion and higher-order information correlation. First, multi-source heterogeneous hypergraphs (MHC) are generated by combining distance, behavior, attribute, and prediction hyperedges. We mine associations among multi-source information, enhancing high-order semantic connections among users and items. Then, a Spatial Density Hypergraph Attention Network (SD-HGAT) is proposed based on composite hyperedges, which enriches the user and item embedding representations by focusing on nodes and hyperedges density. Finally, we design a multiple cross view contrastive learning (MCC) that compares views centered around knowledge graphs with hypergraphs, improving the accuracy of multivariate relationship modeling and enabling multi-level user profiling construction. It is observed that MHCLR outperforms the baselines in terms of Recall, Precision, and NDCG based on the experimental results by Yelp, Last-FM, and Douban.

Multiple hypergraph convolutional network social recommendation using dual contrastive learning

Dual Homogeneity Hypergraph Motifs with Cross-view Contrastive Learning for Multiple Social Recommendations

Inhomogeneous Social Recommendation with Hypergraph Convolutional Networks

Multi-view graph contrastive learning for social recommendation

Dual Variational Graph Reconstruction Learning for Social Recommendation

A Recommendation Method Based on Multi-Source Heterogeneous Hypergraphs and Contrastive Learning

Dual Graph Attention Networks for Deep Latent Representation of Multifaceted Social Effects in Recommender Systems

Disentangled Hypergraph Collaborative Filtering for Social Recommendation

Dual-Channel Multiplex Graph Neural Networks for Recommendation

Multi-Behavior Enhanced Heterogeneous Graph Convolutional Networks Recommendation Algorithm based on Feature-Interaction

Graph-Augmentation-Free Self-Supervised Learning for Social Recommendation

Hypergraph Contrastive Collaborative Filtering

SSGCL: Simple Social Recommendation with Graph Contrastive Learning

Dual Channel Hypergraph Collaborative Filtering

Disentangled Modeling of Social Homophily and Influence for Social Recommendation

SocialGCN: an Efficient Graph Convolutional Network Based Model for Social Recommendation

Multi-Graph Heterogeneous Interaction Fusion for Social Recommendation

Bilateral Sequential Hypergraph Convolution Network for Reciprocal Recommendation

Graph Heterogeneous Multi-Relational Recommendation

Contrastive Graph Learning for Social Recommendation

MCSRec: Modeling Cognitive Similarity in Sequential Recommendation with Social Networks